Smart Lighting Control with Object Detection

An intelligent lighting control system that uses YOLOv8 object detection to automatically adjust light brightness based on detected objects in a camera stream. The system can distinguish between people (triggers lights) and smaller objects like birds (ignored).

Features

• 🎯 Real-time Object Detection using YOLOv8
• 📹 Multiple Camera Sources support (webcam, IP cameras, RTSP, HTTP streams)
• 💡 Intelligent Light Control based on object size and type
• 🔧 FastAPI REST API for control and monitoring
• ⚙️ Configurable Thresholds for object filtering
• 🎚️ Smooth Transitions with debouncing to prevent flickering
• 📊 Real-time Monitoring via API endpoints

Technology Stack

• Object Detection: YOLOv8 (Ultralytics)
• API Framework: FastAPI
• Computer Vision: OpenCV
• Configuration: YAML

Installation

Option 1: Using Docker (Recommended for Team/Production)

Prerequisites:

• Docker Desktop installed (Download)

For Team Members - Pull Pre-built Image:

# Pull the latest image
docker pull rs735my/smart-lighting-detection:latest

# Run with docker-compose
docker-compose up

The application will be available at http://localhost:8000

Note: USB webcams don't work in Docker on Windows. Use IP camera, RTSP stream, or video file. Edit config.yaml:

camera:
  source: "http://192.168.1.100:8080/video"  # IP Webcam from phone
  # OR
  source: "/app/videos/sample.mp4"  # Video file

For Developers - Build Locally:

# Build and run
docker-compose up --build

# Or build without running
docker build -t smart-lighting-detection .

Option 2: Native Python Installation (For USB Webcam Development)

1. Clone or navigate to the project directory

2. Create a virtual environment:

   python -m venv venv
   .\venv\Scripts\Activate.ps1

3. Install dependencies:

   pip install -r requirements.txt

4. Set camera source in config.yaml:

   camera:
     source: "0"  # USB webcam works with native Python

Configuration

Edit config.yaml to customize:

• Camera source (webcam, IP camera URL, video file)
• Detection settings (model type, confidence threshold, target classes)
• Object size filtering (min/max bounding box area)
• Light control (brightness levels, fade duration, auto-off delay)
• Lighting backend (simulated, MQTT, HTTP, Philips Hue)

Usage

Start the Application

python main.py

The API will be available at: http://localhost:8000

API Endpoints

• GET / - API information
• GET /status - Get current system status
• POST /start - Start video processing
• POST /stop - Stop video processing
• GET /stream - Live video stream with detections
• POST /config - Update configuration
• GET /lights/status - Get current light status
• POST /lights/manual - Manually control lights

Interactive API Documentation

Visit http://localhost:8000/docs for Swagger UI documentation.

How It Works

1. Video Capture: Fetches frames from configured camera source
2. Object Detection: YOLOv8 analyzes each frame and detects objects
3. Filtering: Filters objects by class (person) and size (area > threshold)
4. Light Control: Adjusts light brightness based on detected objects
5. Debouncing: Prevents rapid on/off switching with configurable delays

Configuration Examples

Use Webcam

camera:
  source: "0"

Use IP Camera (RTSP)

camera:
  source: "rtsp://username:password@192.168.1.100:554/stream"

Use HTTP Stream

camera:
  source: "http://192.168.1.100:8080/video"

Adjust Detection Sensitivity

detection:
  confidence: 0.6  # Higher = more strict
  min_object_size: 10000  # Larger = only bigger objects

Light Control Modes

Simulated (Testing)

lighting:
  mode: "simulated"

MQTT

lighting:
  mode: "mqtt"
  mqtt:
    broker: "192.168.1.100"
    topic: "home/lights/control"

HTTP API

lighting:
  mode: "http"
  http:
    url: "http://your-light-api.com/control"

Troubleshooting

Camera not connecting

• Verify camera URL/source in config.yaml
• Check network connectivity
• Test with webcam (source: "0") first

No objects detected

• Lower confidence threshold in config
• Verify camera has clear view
• Check target_classes includes "person"

Lights not responding

• Verify lighting.mode is configured correctly
• Check connection to light control system
• Test with mode: "simulated" first

Development

Updating Docker Image (For Maintainers)

When you add new dependencies or make changes:

1. Update requirements.txt with new packages:

   pip freeze > requirements.txt

2. Rebuild the Docker image:

   docker-compose build --no-cache

3. Tag the new version:

   docker tag tmovprojekt-object-detection:latest rs735my/smart-lighting-detection:latest
   # Or with version number
   docker tag tmovprojekt-object-detection:latest rs735my/smart-lighting-detection:v1.1

4. Push to Docker Hub:

   # Login first (if not already)
   docker login
   
   # Push latest
   docker push rs735my/smart-lighting-detection:latest
   
   # Also push versioned tag (recommended)
   docker push rs735my/smart-lighting-detection:v1.1

5. Notify team members to pull the new image:

   docker-compose pull
   docker-compose up

Adding New Python Packages

1. Install package locally:

   pip install package-name

2. Update requirements.txt:

   pip freeze > requirements.txt

3. Rebuild and push Docker image (steps above)

Project Structure

.
├── main.py              # FastAPI application entry point
├── detector.py          # YOLOv8 object detection
├── camera.py            # Camera stream handling
├── light_controller.py  # Light control logic
├── video_processor.py   # Video processing service
├── config.yaml          # Configuration file
├── requirements.txt     # Python dependencies
└── README.md           # This file

License

MIT License

Future Enhancements

• Web dashboard for monitoring
• Multiple camera support
• Zone-based detection
• Activity logging and analytics
• Mobile app integration
• Advanced scheduling rules